Background and significance. Diagnostic errors are highly prevalent, affecting 12 million US adults per year (~1 in 20) in outpatient settings alone. Half are estimated to be harmful, with an estimated 40,000- 80,000 people dying every year in the US because of diagnostic errors. Furthermore, Veterans, who have more medical conditions than non-Veterans do, may be disproportionately affected by diagnostic errors. Indeed, at least 1 million Veterans may have diagnostic errors each year, preventing them from receiving the timely and helpful treatments they deserve. Given a high prevalence of diagnostic errors, researchers have begun to identify the origins of these errors and have attributed cognitive causes to a majority of them. Many times, however, the source of each error has been identified as a cognitive bias, which has been found to be very difficult to detect and address. However, evidence suggests that cognitive characteristics of physicians (e.g., their situation awareness [SA; ability to assess the current situation] and their metacognitive calibration [ability to accurately assess their performance]) and the way they use the electronic health record (EHR) may be two important, yet understudied factors contributing to diagnostic error. These areas of research and an educational intervention to improve such factors to decrease error are the focus of this proposal. This research addresses the overall goal of high quality and safe care for Veterans and the use of health care informatics, a VA HSR&D cross-cutting priority area, by understanding how physicians utilize the EHR to diagnose patients and how we can improve EHR use to improve diagnosis. Research plan. In this proposal, both cognitive characteristics of physicians and patterns of EHR use will be examined as they relate to diagnostic accuracy. Then, an educational intervention aimed at improving these factors will be developed and pilot-tested for the long-term goal of improving diagnostic accuracy and reducing diagnostic errors in Veterans. The specific aims of this research are to: Aim 1) examine the relationship between diagnostic accuracy and cognitive characteristics of physicians in a series of general medical vignettes, Aim 2) investigate patterns of EHR use during diagnostic decision making and related accuracy in a simulated, naturalistic EHR setting using standardized patients, and Aim 3) develop and pilot an educational intervention that provides assessment and feedback on diagnosis-related performance in a naturalistic EHR environment. We will use the SA in Adaptive Decision Making Framework from the human factors field to guide this work. Aim 1 will consist of measuring physicians' cognitive characteristics, including SA and metacognitive calibration obtained while physicians solve validated patient vignettes. Then the relationships between these characteristics and diagnostic accuracy on the vignettes will be examined. Aim 2 will utilize simulation and Naturalistic Decision Making (NDM) methods to examine how physicians utilize the EHR as they diagnose patient vignettes in a simulated EHR environment using standardized patients. These patterns of EHR use will be related to diagnostic accuracy, SA, and metacognitive calibration. Aim 3 will consist of the creation and piloting of an educational intervention aimed at improving factors related to diagnostic accuracy. Career plan. This multidisciplinary research, along with career development and mentoring plans to develop expertise in clinical diagnosis, clinical informatics, systems engineering/human factors, and educational interventions; will increase knowledge of diagnostic errors and enhance my ability to transition to an independent Veterans Affairs (VA) health services researcher.